Fabian Isensee
Fabian Isensee
Hey @beantr96, this is because the patch size must be a multiple of 2**num_pool. So if we want to pool 4 times along that axis, we cannot have 18 because...
Use `nnUNet_determine_postprocessing` :-)
Can you please give some details? Dataset? Configuration? What EXACTLY did you run/try to do?
You issue really makes little sense to me. You have not given any details on what the problem is. Does the segmentation not work? Are the metrics not what you...
OK so you run inference twice and get two different results? That is strange and should not happen, but I also do not have an explanation for this. There is...
You need to remove the unused labels from your segmentation images
(remove = replace with 0)
Hi, I am a bit short in time now. It's best to start in the nnUNetTrainerV2.initialize function where the data augmentation is initialized. From there you can work your way...
I means that there are values in your segmentation image that don't belong there! Your segmentation must be an integer image with 0=background, 1=class1, 2=class2 etc
Hi, thanks for pointing this out. Have you verified on one of the standard datasets that this happens there as well? If it works fine on those we might have...